Method and apparatus for determining response for user input data, and medium

Patent metadata

Field	Value
Publication number	US-11238050-B2
Application number	US-201916459282-A
Country	US
Kind code	B2
Filing date	Jul 1, 2019
Priority date	Jul 3, 2018
Publication date	Feb 1, 2022
Grant date	Feb 1, 2022

Abstract

According to exemplary embodiments of the present disclosure, a method and an apparatus for determining a response for user input data, a device, and a computer-readable storage medium are provided. The method includes: obtaining search results associated with the user input data. The method also includes: determining correlation degrees between the search results and the user input data. The method also includes: building candidate content for the user input data based on the search results and the correlation degrees. The candidate content includes a paragraph associated with the user input data in the search results. The method also includes: determining the response for the user input data from the candidate content.

First claim

What is claimed is:

1. A method for determining a response for user input data, performed by an intelligent question answering device, comprising:
obtaining search results associated with the user input data by using a search engine through a network;
determining correlation degrees between the search results and the user input data based on a keyword in the user input data;
building candidate content for the user input data based on the search results and the correlation degrees, comprising: selecting a predetermined number of search results from the search results according to the correlation degrees between the search results and the user input data; determining correlation degrees between paragraphs in the predetermined number of search results selected and the user input data; and generating the candidate content for the user input data based on at least one paragraph selected from the paragraphs according to the correlation degrees between the paragraphs in the predetermined number of search results selected and the user input data; and
determining the response for the user input data from segments of the candidate content based on probabilities of the user input data in the segments, comprising: selecting a segment in a descending order of possibilities, and determining the response for the user input data based on the segment selected.

2. The method of claim 1 , wherein, determining the correlation degrees between the search results and the user input data comprises:
determining the correlation degrees between the search results and the user input data based on historical browsing behavior information associated with the keyword.

3. The method of claim 1 , wherein selecting the predetermined number of search results from the search results according to the correlation degrees between the search results and the user input data comprises:
selecting the predetermined number of search results from the search results in a descending order of the correlation degrees between the search results and the user input data.

4. The method of claim 1 , wherein the at least one paragraph is selected from the paragraphs in the predetermined number of search results in a descending order of the correlation degrees between the paragraphs in the predetermined number of search results selected and the user input data.

5. The method of claim 1 , wherein, determining the correlation degrees between the paragraphs in the predetermined number of search results selected and the user input data comprises:
determining correlation degrees between the paragraphs and the keyword as the correlation degrees between the paragraphs in the predetermined number of search results selected and the user input data.

6. The method of claim 1 , wherein, determining the response for the user input data based on the segment selected comprises:
applying the segment selected and the user input data to a response generation model, to generate the response, in which, the response generation mode is determined based on training the user input data and training the segments.

7. The method of claim 1 , further comprising:
providing the response to a user in at least one form of: speech, text and table.

8. An apparatus for determining a response for user input data, applied to an intelligent question answering device, comprising:
one or more processors and a memory; wherein the one or more processors are configured to:
obtain search results associated with the user input data by using a search engine through a network;
determine correlation degrees between the search results and the user input data based on a keyword in the user input data;
build candidate content for the user input data based on the search results and the correlation degrees by performing acts of: selecting a predetermined number of search results from the search results according to the correlation degrees between the search results and the user input data; determining correlation degrees between paragraphs in the predetermined number of search results selected and the user input data; and generating the candidate content for the user input data based on at least one paragraph selected from the paragraphs according to the correlation degrees between the paragraphs in the predetermined number of search results selected and the user input data; and
determine the response for the user input data from segments of the candidate content based on probabilities of the user input data in the segments, comprising: selecting a segment in a descending order of possibilities, and determining the response for the user input data based on the segment selected.

9. The apparatus of claim 8 , wherein, the one or more processors determine the correlation degrees between the search results and the user input data by performing an act of:
determining the correlation degrees between the search results and the user input data based on historical browsing behavior information associated with the keyword.

10. The apparatus of claim 8 , wherein, the one or more processors select the predetermined number of search results from the search results according to the correlation degrees between the search results and the user input data by performing an act of: selecting the predetermined number of search results from the search results in a descending order of the correlation degrees between the search results and the user input data; and
the at least one paragraph is selected from the paragraphs in the predetermined number of search results in a descending order of the correlation degrees between the paragraphs in the predetermined number of search results selected and the user input data.

11. The apparatus of claim 8 , wherein, the one or more processors determine the correlation degrees between the paragraphs in the predetermined number of search results selected and the user input data by performing an act of:
determining correlation degrees between the paragraphs and the keyword as the correlation degrees between the paragraphs in the predetermined number of search results selected and the user input data.

12. The apparatus of claim 8 , wherein, the one or more processors determine the response for the user input data based on the segment selected by performing an act of:
applying the segment selected and the user input data to a response generation mode, to generate the response, in which, the response generation mode is determined based on training the user input data and training the segments.

13. The apparatus of claim 8 , wherein the one or more processors are further configured to:
provide the response to a user in at least one form of: speech, text and table.

14. A non-transitory computer readable storage medium having a computer program stored thereon, wherein, when the computer program is executed by a processor, the processor is caused to perform a method for determining a response for user input data, wherein the method is performed by an intelligent question answering device and comprises:
obtaining search results associated with the user input data by using a search engine through a network;
determining correlation degrees between the search results and the user input data based on a keyword in the user input data;
building candidate content for the user input data based on the search results and the correlation degrees, comprising: selecting a predetermined number of search results from the search results according to the correlation degrees between the search results and the user input data; determining correlation degrees between paragraphs in the predetermined number of search results selected and t

Assignees

• Baidu online network technology beijing co ltd

Inventors

• Fan Miao
• Feng Chao
• Sun Mingming
• Li Ping
• Wang Haifeng

Classifications

• G06F16/248

  Presentation of query results · CPC title

• G06F16/24573 Primary

  using data annotations, e.g. user-defined metadata · CPC title

• G06F16/3329 Primary

  Natural language query formulation · CPC title

• G06F16/24578

  using ranking · CPC title